Kamra PMS
Try Kamra — the open-source, AI-native hotel PMS. Front desk, direct booking, housekeeping, folios and GST billing. One-tap demo logins for every role.
翻譯準備中
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
Studio software for photographers: booking, proofing galleries, contracts, invoices and Stripe payments, with AI inquiry triage, gallery search, culling and a Studio Assistant. Live at photoezcloud.com.
翻譯準備中
正在載入完整 README。
檢視完整 README ↗AI-assisted studio software for photographers, live at photoezcloud.com. Booking, proofing galleries, contracts, invoices and payments in one place, with AI that does the busywork: sorting inquiries and drafting replies, finding the right shots, flagging the ones to cut, and answering questions about the studio.
PhotoEZ Cloud is the SaaS successor to the PhotoEZ WordPress plugin suite, which I built and sell through EJS Tech. The plugins give photographers galleries, booking, contracts, and invoicing on their own sites. This project rebuilds the whole workflow as a hosted app and puts AI features at the center.
Live at photoezcloud.com. Photographers can sign up, take real payments through Stripe, and deliver galleries today; it runs on a VPS with zero-downtime deploys and nightly off-site backups. Free plan, and a 14-day Pro trial with no card.
Each one has an eval built from real studio work, run before any change to the prompt or model ships. The evals live in web/evals.
| Feature | What it does | How it works | Measured |
|---|---|---|---|
| Inquiry triage | Reads a new inquiry, pulls out the session type, date, budget and questions, flags what needs the photographer, and drafts a reply | Claude with structured outputs; replies point to the booking page and FAQ instead of quoting prices | Reply safe to send: 88% → 96%; never missed an inquiry that needed the photographer (eval) |
| Gallery search | Finds photos from plain words like "first dance" or "girl in a blue dress" | Claude Haiku tags each photo with everyday words; whole-word, phrase-aware Postgres search | No wrong matches: 70% → 93% (eval) |
| Culling help | Flags blurry frames, too dark/bright, missed flash, closed eyes, and near-duplicates before the client sees them | Runs in the photographer's browser at no server cost: tile sharpness, exposure percentiles (so low-key portraits and darker skin tones aren't flagged), difference hashes, and a face-landmark model | Tested on real galleries |
| Studio Assistant | Answers "which galleries expire this week?" and acts on it, like drafting reminder emails | Claude tool use; every action waits on the photographer's Approve card, checked in code, with full step traces | No made-up info: 96% (eval, architecture) |
Every AI call is recorded with its token cost, so each studio's AI use is tracked against its plan's allowance (cost tracking).
Built by Elle Jones, Full-Stack AI Engineer in Portland, OR, and founder of EJS Tech. Elle is also a working photographer, which is where these features come from.